For globally searching the impressive GRETIL mass of e-texts kindly provided to us by Reinhold Grünendahl for personal use, a student of mine recommends the Jedit software (about which, see below). 
Best wishes,
Christophe Vielle

Début du message réexpédié :

Pour le moteur de recherche, on peut simplement utiliser l'éditeur (multi-plateforme et
opensource) Jedit [*], facile à trouver et à installer. Il permet la
recherche dans un dossier complet. Le résultat se présente avec nombre
d'occurrences, nom du fichier, numéro de la ligne, kwik (key word in context). Ci-joint en
image la recherche du mot "upaniṣad" dans tout le corpus de 404 millions
de caractères, et cela prend environ une minute. C'est moins rapide
qu'en Perl (le langage fait pour la recherche
documentaire, extrêmement rapide), mais utilisable.



* About Jedit, Wujastyk wrote on Indology list in October 2002: 

The following may not be the full answer, but colleagues tell me that
Jedit (from www.jedit.com) is very good at suporting Unicode.  My friend
Gerhard Brey has used Jedit successfully to create XML files containing
Unicode text of both Roman and Arabic materials.  He tells me that Jedit
is better than Yudit as a Unicode editor.

I've installed Jedit under Windows 2000, just to play with - it was very
easy indeed.  Being a Java program, Jedit will run on any platform, Mac,
Linux, etc., and behave in essentially the same way.  If you load a
Unicode font (there are now quite a few around) as the default screen
font, you get a nice system, and it's portable and free.

Jedit can be used for writing TeX files, or anything else.  It has
built-in syntax highlighting for TeX and many other structured languages
such as XML, HTML, programming languages, etc.  And it's extensible, and
apparently has quite a bit of external support from an enthusiastic
community.
–––––––––––––––––––
Louvain-la-Neuve